AI can help identify women at risk of serious cardiovascular disease by analyzing routine mammography scans. That is the conclusion of a large-scale study. The research shows that AI algorithms can detect and quantify calcium deposits in the arteries of the breast, an indicator associated with cardiovascular disease, using standard X-ray mammograms already performed in breast cancer screening programs.
The approach could offer an important opportunity to detect cardiovascular risk in women earlier, addressing a long-standing gap in diagnosis and treatment. The study was published in the European Heart Journal.
Underdiagnosed
Cardiovascular disease is the leading cause of death among women worldwide. Yet research consistently shows that women are less frequently diagnosed and treated compared to men. According to Dr. Hari Trivedi of Emory University in Atlanta, who led the study, this disparity highlights the need for innovative screening strategies.
“Mammograms are already widely used for breast cancer screening,” he explains. “These images can also show calcium deposits in breast arteries, which are associated with cardiovascular disease. We wanted to investigate whether AI could analyze these images to identify women at risk, without requiring additional tests, costs, or inconvenience.”
Large-scale analysis
The study included 123,762 women who had undergone breast cancer screening and had no known cardiovascular disease at the time of imaging. Researchers used an AI model to assess breast arterial calcification (BAC), calcium deposits in the arteries within breast tissue visible on mammograms.
BAC is considered a marker of arterial hardening and is associated with increased risk of cardiovascular conditions such as heart attack, stroke, heart failure, and cardiovascular mortality. The AI system classified the degree of calcification into four categories:
- Absent
- Mild
- Moderate
- Severe
Researchers then compared these classifications with long-term health outcomes to determine whether BAC predicted cardiovascular events.
Clear correlation with cardiovascular outcomes
The results revealed a strong association between the amount of calcification and future cardiovascular risk. Women with mild calcification were approximately 30 percent more likely to develop serious cardiovascular disease compared with women without calcification. For women with moderate calcification, the risk was more than 70 percent higher. Those with severe calcification faced a risk two to three times greater.
Importantly, the association remained significant even after adjusting for established risk factors such as diabetes and smoking. The findings were also consistent in women younger than 50, a group often considered to have relatively low cardiovascular risk.
“This is the largest study of its kind,” says Trivedi. “It includes women of multiple ethnic backgrounds across two major U.S. health systems. The more calcium visible in breast arteries on a mammogram, the higher the risk of serious cardiovascular events.”
A dual diagnostic tool
The researchers emphasize that integrating AI-based BAC analysis into existing mammography workflows could enable large-scale cardiovascular risk screening without additional infrastructure.
For patients, this could mean that a routine mammogram provides insights beyond cancer detection. If calcification is detected, clinicians could initiate preventive measures such as cholesterol testing, lifestyle counseling, or medication.
For healthcare providers, the technology offers a practical method to identify women who may otherwise remain unaware of their cardiovascular risk.
Opportunity for preventive care
In an accompanying editorial, Professor Lori B. Daniels of the University of California, San Diego, highlights the potential public health impact of the approach.
Screening participation for breast cancer is already high. In the European Union, about two-thirds of women aged 50–69 report having had a mammogram in the previous two years. In the United States, nearly 70 percent of women aged 45 and older are up to date with mammography screening according to American Cancer Society guidelines.
By contrast, fewer than 40 percent of women report knowing their cholesterol levels. Breast arterial calcification could therefore help close this prevention gap. “It leverages a screening platform that millions of women already use,” Daniels notes, “to identify cardiovascular risk in those who might otherwise not engage with preventive care.” The editorial concludes that the evidence supporting AI-quantified BAC is now strong enough to move toward clinical implementation.
Clinical integration
The research team is currently preparing a clinical trial to evaluate how AI-based BAC detection can be integrated into clinical workflows. Key considerations include embedding the AI tool within imaging systems and establishing guidelines for reporting results to clinicians and patients.
If successfully implemented, this approach could transform routine mammography into a dual-purpose screening tool, simultaneously detecting breast cancer and identifying women at risk of cardiovascular disease.
Given that cardiovascular disease remains the leading cause of death among women worldwide, the opportunity to use an existing screening moment to support prevention may prove highly valuable for population health.
Comparable research
Earlier this year a comparable research from Penn State College of Medicine also suggested that calcium deposits in breast arteries (BAC) visible on mammograms can help predict a woman’s future risk of cardiovascular disease. BAC appears in about 15–25% of screening mammograms but is usually not reported because it is unrelated to breast cancer.
Using AI to analyze mammograms from over 10,000 women, researchers found that the presence and progression of BAC strongly correlated with higher risks of heart attack, stroke, heart failure and mortality. Women whose calcification worsened had up to twice the risk of major cardiovascular events. Because BAC can be detected in existing mammography images, this approach could enable earlier cardiovascular risk identification without additional tests, radiation or healthcare costs.
